tools/3_visualise_transcribed_teams.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook visualises the transcribed teams among the other teams in the feature space spanned by task performance and learning outcome, as well as the distribution of their number of attempts and turns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import re\n",
    "import pickle\n",
    "import datetime\n",
    "\n",
    "import pathlib as pl\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "import seaborn as sns\n",
    "\n",
    "from read_utils import get_team_no_from_transcript_filename\n",
    "\n",
    "sns.set_theme(\n",
    "    context='paper',\n",
    "    font_scale=1.5,\n",
    "    rc={'text.usetex': True}\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define paths."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Inputs.\n",
    "data_dir = pl.Path('../data')\n",
    "\n",
    "transcripts_dir = data_dir.joinpath('transcripts')\n",
    "\n",
    "processed_data_dir = pl.Path('../processed_data')\n",
    "log_features_file = processed_data_dir.joinpath(\n",
    "    'log_features/justhink19_log_features_task_level.csv')\n",
    "learning_file = processed_data_dir.joinpath(\n",
    "    'learning_features/justhink19_learning_features.csv')\n",
    "\n",
    "# Outputs.\n",
    "plots_dir = pl.Path('../figures')\n",
    "perf_vs_learning_file = plots_dir.joinpath('performance_vs_learning.png')\n",
    "\n",
    "# Create output directories if they do not exist.\n",
    "for folder in [plots_dir]:\n",
    "    if not folder.exists():\n",
    "        folder.mkdir(parents=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "code_folding": []
   },
   "source": [
    "## Load data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read learning data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>A_RLG</th>\n",
       "      <th>B_RLG</th>\n",
       "      <th>RLG</th>\n",
       "      <th>A_pretest</th>\n",
       "      <th>A_posttest</th>\n",
       "      <th>B_pretest</th>\n",
       "      <th>B_posttest</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>team_no</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>-0.200000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.100000</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.285714</td>\n",
       "      <td>0.142857</td>\n",
       "      <td>0.7</td>\n",
       "      <td>0.7</td>\n",
       "      <td>0.3</td>\n",
       "      <td>0.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.200000</td>\n",
       "      <td>-0.375000</td>\n",
       "      <td>-0.087500</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.8</td>\n",
       "      <td>0.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.200000</td>\n",
       "      <td>0.250000</td>\n",
       "      <td>0.225000</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.400000</td>\n",
       "      <td>0.366667</td>\n",
       "      <td>0.4</td>\n",
       "      <td>0.6</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            A_RLG     B_RLG       RLG  A_pretest  A_posttest  B_pretest  \\\n",
       "team_no                                                                   \n",
       "6       -0.200000  0.000000 -0.100000        0.5         0.4        0.6   \n",
       "7        0.000000  0.285714  0.142857        0.7         0.7        0.3   \n",
       "8        0.200000 -0.375000 -0.087500        0.5         0.6        0.8   \n",
       "9        0.200000  0.250000  0.225000        0.5         0.6        0.6   \n",
       "10       0.333333  0.400000  0.366667        0.4         0.6        0.5   \n",
       "\n",
       "         B_posttest  \n",
       "team_no              \n",
       "6               0.6  \n",
       "7               0.5  \n",
       "8               0.5  \n",
       "9               0.7  \n",
       "10              0.7  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learning_df = pd.read_csv(learning_file, index_col='team_no')\n",
    "\n",
    "learning_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read task features data (at the team level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>corpus_no</th>\n",
       "      <th>last_error</th>\n",
       "      <th>max_error</th>\n",
       "      <th>min_error</th>\n",
       "      <th>range_error</th>\n",
       "      <th>avg_error</th>\n",
       "      <th>std_error</th>\n",
       "      <th>noninc_error</th>\n",
       "      <th>slope_error</th>\n",
       "      <th>success</th>\n",
       "      <th>...</th>\n",
       "      <th>avg_action_per_turn</th>\n",
       "      <th>num_turn</th>\n",
       "      <th>num_action</th>\n",
       "      <th>duration</th>\n",
       "      <th>num_edit</th>\n",
       "      <th>freq_add</th>\n",
       "      <th>freq_histload</th>\n",
       "      <th>num_submitselect</th>\n",
       "      <th>num_histview</th>\n",
       "      <th>num_help</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>team_no</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.090909</td>\n",
       "      <td>0.045455</td>\n",
       "      <td>0.045455</td>\n",
       "      <td>True</td>\n",
       "      <td>-11.000000</td>\n",
       "      <td>True</td>\n",
       "      <td>...</td>\n",
       "      <td>3.571429</td>\n",
       "      <td>28</td>\n",
       "      <td>100</td>\n",
       "      <td>1350.461</td>\n",
       "      <td>56</td>\n",
       "      <td>0.660714</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>9</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>1.909091</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>1.772727</td>\n",
       "      <td>0.505051</td>\n",
       "      <td>0.505908</td>\n",
       "      <td>False</td>\n",
       "      <td>-6.363863</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>4.285714</td>\n",
       "      <td>98</td>\n",
       "      <td>420</td>\n",
       "      <td>1581.128</td>\n",
       "      <td>206</td>\n",
       "      <td>0.922330</td>\n",
       "      <td>0.208333</td>\n",
       "      <td>9</td>\n",
       "      <td>24</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1</td>\n",
       "      <td>0.409091</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.308081</td>\n",
       "      <td>0.114948</td>\n",
       "      <td>False</td>\n",
       "      <td>9.555985</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>5.057692</td>\n",
       "      <td>52</td>\n",
       "      <td>263</td>\n",
       "      <td>1559.334</td>\n",
       "      <td>107</td>\n",
       "      <td>0.897196</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>28</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1</td>\n",
       "      <td>0.318182</td>\n",
       "      <td>0.363636</td>\n",
       "      <td>0.045455</td>\n",
       "      <td>0.318182</td>\n",
       "      <td>0.234848</td>\n",
       "      <td>0.103041</td>\n",
       "      <td>False</td>\n",
       "      <td>-20.335135</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>7.823529</td>\n",
       "      <td>34</td>\n",
       "      <td>266</td>\n",
       "      <td>2152.799</td>\n",
       "      <td>68</td>\n",
       "      <td>0.720588</td>\n",
       "      <td>0.645161</td>\n",
       "      <td>0</td>\n",
       "      <td>31</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1</td>\n",
       "      <td>0.454545</td>\n",
       "      <td>0.681818</td>\n",
       "      <td>0.136364</td>\n",
       "      <td>0.545455</td>\n",
       "      <td>0.318182</td>\n",
       "      <td>0.202430</td>\n",
       "      <td>False</td>\n",
       "      <td>-8.873950</td>\n",
       "      <td>False</td>\n",
       "      <td>...</td>\n",
       "      <td>5.882353</td>\n",
       "      <td>51</td>\n",
       "      <td>300</td>\n",
       "      <td>1375.537</td>\n",
       "      <td>108</td>\n",
       "      <td>0.981481</td>\n",
       "      <td>0.275862</td>\n",
       "      <td>1</td>\n",
       "      <td>29</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         corpus_no  last_error  max_error  min_error  range_error  avg_error  \\\n",
       "team_no                                                                        \n",
       "6                1    0.000000   0.090909   0.000000     0.090909   0.045455   \n",
       "7                1    0.136364   1.909091   0.136364     1.772727   0.505051   \n",
       "8                1    0.409091   0.500000   0.136364     0.363636   0.308081   \n",
       "9                1    0.318182   0.363636   0.045455     0.318182   0.234848   \n",
       "10               1    0.454545   0.681818   0.136364     0.545455   0.318182   \n",
       "\n",
       "         std_error  noninc_error  slope_error  success  ...  \\\n",
       "team_no                                                 ...   \n",
       "6         0.045455          True   -11.000000     True  ...   \n",
       "7         0.505908         False    -6.363863    False  ...   \n",
       "8         0.114948         False     9.555985    False  ...   \n",
       "9         0.103041         False   -20.335135    False  ...   \n",
       "10        0.202430         False    -8.873950    False  ...   \n",
       "\n",
       "         avg_action_per_turn  num_turn  num_action  duration  num_edit  \\\n",
       "team_no                                                                  \n",
       "6                   3.571429        28         100  1350.461        56   \n",
       "7                   4.285714        98         420  1581.128       206   \n",
       "8                   5.057692        52         263  1559.334       107   \n",
       "9                   7.823529        34         266  2152.799        68   \n",
       "10                  5.882353        51         300  1375.537       108   \n",
       "\n",
       "         freq_add  freq_histload  num_submitselect  num_histview  num_help  \n",
       "team_no                                                                     \n",
       "6        0.660714       0.000000                 9             5         3  \n",
       "7        0.922330       0.208333                 9            24         8  \n",
       "8        0.897196       0.000000                 0            28        14  \n",
       "9        0.720588       0.645161                 0            31         6  \n",
       "10       0.981481       0.275862                 1            29        21  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "task_df = pd.read_csv(log_features_file, index_col='team_no')\n",
    "\n",
    "task_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find dialogue transcripts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{7: PosixPath('../data/transcripts/justhink19_transcript_07.csv'),\n",
       " 8: PosixPath('../data/transcripts/justhink19_transcript_08.csv'),\n",
       " 9: PosixPath('../data/transcripts/justhink19_transcript_09.csv'),\n",
       " 10: PosixPath('../data/transcripts/justhink19_transcript_10.csv'),\n",
       " 11: PosixPath('../data/transcripts/justhink19_transcript_11.csv'),\n",
       " 17: PosixPath('../data/transcripts/justhink19_transcript_17.csv'),\n",
       " 18: PosixPath('../data/transcripts/justhink19_transcript_18.csv'),\n",
       " 20: PosixPath('../data/transcripts/justhink19_transcript_20.csv'),\n",
       " 28: PosixPath('../data/transcripts/justhink19_transcript_28.csv'),\n",
       " 47: PosixPath('../data/transcripts/justhink19_transcript_47.csv')}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transcript_files = sorted([f for f in transcripts_dir.glob('*.csv')])\n",
    "transcript_files = {get_team_no_from_transcript_filename(f):f for f in transcript_files}\n",
    "\n",
    "transcript_files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[7, 8, 9, 10, 11, 17, 18, 20, 28, 47]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transcribed_teams = sorted(transcript_files.keys())\n",
    "transcribed_teams"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate team list that have audio recordings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Audio recordings for 38 out of 39 teams exist.\n"
     ]
    }
   ],
   "source": [
    "with_audio_teams = list(learning_df.index)\n",
    "\n",
    "# We do not have complete audio recordings for team 33.\n",
    "with_audio_teams.remove(33) \n",
    "\n",
    "print('Audio recordings for {} out of {} teams exist.'.format(\n",
    "    len(with_audio_teams), len(learning_df)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Print the total task duration in hours for all data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "15.926999166666665"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "values = [td / 60 / 60 for td in task_df['duration']]\n",
    "sum(values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualise teams in learning vs. performance space."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Make the input table for the plot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>min_error</th>\n",
       "      <th>RLG</th>\n",
       "      <th>kind</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>team_no</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-10.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>14.2857</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>-8.7500</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>22.5000</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>36.6667</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>36.6667</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>25.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>-14.2857</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>22.727273</td>\n",
       "      <td>-2.0833</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>10.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>25.0000</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18.181818</td>\n",
       "      <td>-56.2500</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-8.3333</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-52.0833</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>10.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>16.6667</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>50.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>22.727273</td>\n",
       "      <td>3.7500</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>22.727273</td>\n",
       "      <td>16.6667</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>-1.7857</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>12.5000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>-5.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>37.5000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>18.181818</td>\n",
       "      <td>40.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>10.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>4.545455</td>\n",
       "      <td>-11.1111</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>22.727273</td>\n",
       "      <td>1.6667</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>31.818182</td>\n",
       "      <td>-12.5000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>18.181818</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>27.272727</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>9.090909</td>\n",
       "      <td>-8.3333</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>9.090909</td>\n",
       "      <td>5.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>18.181818</td>\n",
       "      <td>25.0000</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>8.3333</td>\n",
       "      <td>other</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>13.636364</td>\n",
       "      <td>-12.6984</td>\n",
       "      <td>transcribed subset</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         min_error      RLG                kind\n",
       "team_no                                        \n",
       "6         0.000000 -10.0000               other\n",
       "7        13.636364  14.2857  transcribed subset\n",
       "8        13.636364  -8.7500  transcribed subset\n",
       "9         4.545455  22.5000  transcribed subset\n",
       "10       13.636364  36.6667  transcribed subset\n",
       "11        4.545455  36.6667  transcribed subset\n",
       "12        4.545455  25.0000               other\n",
       "14       13.636364 -14.2857               other\n",
       "15       22.727273  -2.0833               other\n",
       "16        0.000000  10.0000               other\n",
       "17        0.000000  25.0000  transcribed subset\n",
       "18       18.181818 -56.2500  transcribed subset\n",
       "19        0.000000  -8.3333               other\n",
       "20        0.000000 -52.0833  transcribed subset\n",
       "21        0.000000  10.0000               other\n",
       "22        4.545455   0.0000               other\n",
       "23        4.545455  16.6667               other\n",
       "25        4.545455  50.0000               other\n",
       "26       22.727273   3.7500               other\n",
       "27       22.727273  16.6667               other\n",
       "28        0.000000   0.0000  transcribed subset\n",
       "29        0.000000  -1.7857               other\n",
       "30        4.545455  12.5000               other\n",
       "31        4.545455  -5.0000               other\n",
       "32       13.636364  37.5000               other\n",
       "34       18.181818  40.0000               other\n",
       "35       13.636364   0.0000               other\n",
       "36       13.636364  10.0000               other\n",
       "37        4.545455 -11.1111               other\n",
       "38       22.727273   1.6667               other\n",
       "39       31.818182 -12.5000               other\n",
       "40       18.181818   0.0000               other\n",
       "41       27.272727   0.0000               other\n",
       "42        9.090909  -8.3333               other\n",
       "44        9.090909   5.0000               other\n",
       "45       18.181818  25.0000               other\n",
       "46       13.636364   8.3333               other\n",
       "47       13.636364 -12.6984  transcribed subset"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Put performance metric values on the x axis.\n",
    "perf_col = 'min_error'\n",
    "x_values = task_df[perf_col] \n",
    "\n",
    "# Put learning metric values on the y axis.\n",
    "learn_col = 'RLG'\n",
    "y_values = learning_df[learn_col]\n",
    "\n",
    "\n",
    "# Merge the tables.\n",
    "df = pd.concat([x_values, y_values], axis=1)\n",
    "\n",
    "# Omit the teams that do not have the audio transcripts.\n",
    "df = df.loc[df.index.isin(with_audio_teams)]\n",
    "\n",
    "# Assign a type for each team, depending on if it is transcribed or not.\n",
    "df['kind'] = ['transcribed subset' if t in transcribed_teams else 'other'\n",
    "              for t in df.index]\n",
    "\n",
    "# Scale by 100 to represent percentages.\n",
    "def scale_hundred(x): return 100 * x\n",
    "df.min_error = df.min_error.apply(scale_hundred)\n",
    "df.RLG = df.RLG.apply(scale_hundred)\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate the plot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "code_folding": [],
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot is exported to ../figures/performance_vs_learning.pdf\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_perf_vs_error(col_x,\n",
    "                       col_y,\n",
    "                       col_k,\n",
    "                       df,\n",
    "                       k_is_color=False,\n",
    "                       colors=['#e48066', '#6bacd1'],\n",
    "                       scatter_alpha=.8):\n",
    "\n",
    "    def colored_scatter(x, y, c=None, label=None):\n",
    "        def scatter(*args, **kwargs):\n",
    "            args = (x, y)\n",
    "            if c is not None:\n",
    "                kwargs['c'] = c\n",
    "            kwargs['alpha'] = scatter_alpha\n",
    "            kwargs['label'] = label\n",
    "            plt.scatter(*args, **kwargs)\n",
    "\n",
    "        return scatter\n",
    "\n",
    "    df = df.copy()\n",
    "    col_x = col_x.replace('_', '\\_')\n",
    "    col_y = col_y.replace('_', '\\_')\n",
    "    df.columns = [c.replace('_', '\\_') for c in df.columns]\n",
    "\n",
    "    prev_x_counts = None\n",
    "    prev_y_counts = None\n",
    "\n",
    "    step = 10\n",
    "    g = sns.JointGrid(\n",
    "        x=col_x,\n",
    "        y=col_y,\n",
    "        xlim=[-1, 35 + 1],\n",
    "        ylim=[-60 - step / 2, 60 + step / 2],\n",
    "        data=df,\n",
    "    )\n",
    "\n",
    "    sns.despine(ax=g.ax_joint, left=True)\n",
    "\n",
    "    color = None\n",
    "    legends = []\n",
    "    handles = []\n",
    "    for i, (name, df_group) in enumerate(reversed(tuple(df.groupby(col_k)))):\n",
    "        legends.append(name)\n",
    "        if k_is_color:\n",
    "            color = name\n",
    "\n",
    "        g.plot_joint(colored_scatter(df_group[col_x], df_group[col_y], colors[i],\n",
    "                                     label=name))\n",
    "\n",
    "        # Plot performance.\n",
    "        # Draw a KDE plot.\n",
    "        sns.kdeplot(\n",
    "            df_group[col_x].values,\n",
    "            ax=g.ax_marg_x,\n",
    "            color=colors[i],\n",
    "            shade=True,\n",
    "        )\n",
    "        # Draw the mean line segments.\n",
    "        m = np.mean(df_group[col_x])\n",
    "        g.ax_marg_x.axvline(\n",
    "            m,\n",
    "            ls='--',\n",
    "            lw=4,\n",
    "            c=colors[i],\n",
    "            alpha=0.9,\n",
    "        )\n",
    "\n",
    "        # Plot learning.\n",
    "        # Draw a KDE plot.\n",
    "        sns.kdeplot(\n",
    "            y=df_group[col_y].values,\n",
    "            ax=g.ax_marg_y,\n",
    "            color=colors[i],\n",
    "            shade=True,\n",
    "        )\n",
    "        # Draw the mean line segments.\n",
    "        m = np.mean(df_group[col_y])\n",
    "        g.ax_marg_y.axhline(\n",
    "            m,\n",
    "            ls='--',\n",
    "            lw=4,\n",
    "            c=colors[i],\n",
    "            alpha=0.9,\n",
    "        )\n",
    "\n",
    "    # Adjust the legend.\n",
    "    g.ax_joint.legend(markerscale=2, handletextpad=0.1)\n",
    "    g.ax_joint.xaxis.set_minor_locator(AutoMinorLocator(n=2))\n",
    "    g.ax_joint.yaxis.set_minor_locator(AutoMinorLocator(n=2))\n",
    "\n",
    "    # Adjust the grid.\n",
    "    plt.grid(alpha=0.85)\n",
    "    plt.grid(which='minor', alpha=0.3)\n",
    "\n",
    "    # Adjust the background colors.\n",
    "    g.ax_marg_x.set_facecolor('white')\n",
    "    g.ax_marg_y.set_facecolor('white')\n",
    "\n",
    "    # Add annotations one by one with a loop.\n",
    "    takens = dict()\n",
    "    for i, row in df.iterrows():\n",
    "        if row['kind'] != 'transcribed subset':\n",
    "            continue\n",
    "        offset = 0.8\n",
    "        plt.text(\n",
    "            row[col_x] + offset, row[col_y], str(i),\n",
    "            fontsize='small',\n",
    "        )\n",
    "\n",
    "    # Set axis labels.\n",
    "    # g.ax_joint.set_xlabel(r\"minimum error \\texttt{error} (\\%)\")\n",
    "    # g.ax_joint.set_ylabel(r\"relative learning gain \\texttt{learn} (\\%)\")\n",
    "    g.ax_joint.set_xlabel(r\"\\texttt{error} (\\%)\")\n",
    "    g.ax_joint.set_ylabel(r\"\\texttt{learn} (\\%)\")\n",
    "\n",
    "    \n",
    "    return g\n",
    "\n",
    "\n",
    "g = plot_perf_vs_error(perf_col, learn_col, 'kind', df=df)\n",
    "\n",
    "# Export as vector graphics - pdf.\n",
    "export_file = perf_vs_learning_file.with_suffix('.pdf')\n",
    "plt.savefig(export_file, bbox_inches='tight')\n",
    "print('Plot is exported to {}'.format(export_file))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualise the teams in number of attempts and turns."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_value(team_no, feature_name): \n",
    "    return task_df.loc[team_no][feature_name]\n",
    "\n",
    "\n",
    "def learning_value(team_no, feature_name): \n",
    "    return learning_df.loc[team_no, feature_name]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_hist(f_name, bins=10,\n",
    "              func=feature_value, legend=True, ax=None,\n",
    "              colors=['#6bacd1', '#e48066'],\n",
    "              n_xminor=2, save=False):\n",
    "\n",
    "    values_1 = [func(t, f_name)\n",
    "                for t in with_audio_teams if t not in transcribed_teams]\n",
    "    values_2 = [func(t, f_name) for t in transcribed_teams]\n",
    "    \n",
    "    if ax is None:\n",
    "        fig, ax = plt.subplots(tight_layout=True)\n",
    "    else:\n",
    "        fig = None\n",
    "\n",
    "    labels = ['other', 'transcribed subset']\n",
    "\n",
    "    ax.hist(\n",
    "        [values_1, values_2],  # , values_3],\n",
    "        bins,\n",
    "        histtype='barstacked',\n",
    "        color=colors,\n",
    "        label=labels,\n",
    "        align='mid',\n",
    "        edgecolor='k', alpha=0.5,\n",
    "    )\n",
    "\n",
    "    if legend:\n",
    "        handles, labels = ax.get_legend_handles_labels()\n",
    "        ax.legend(handles[::-1], labels[::-1], framealpha=1)\n",
    "\n",
    "    m1, m2 = np.mean(values_1), np.mean(values_2)\n",
    "    ymin, ymax = 0, 1\n",
    "    ax.axvline(m1, ymax=ymax, c=colors[0], ls='--', lw=6, alpha=0.9)\n",
    "\n",
    "    ymin = 0.05\n",
    "    ax.axvline(m2, ymin=ymin, ymax=ymax, c=colors[1], ls='--', lw=6, alpha=0.9)\n",
    "\n",
    "    ax.grid(alpha=0.9)\n",
    "    ax.grid(which='minor', alpha=0.5)\n",
    "\n",
    "    ax.set_ylabel('team count')\n",
    "    ax.xaxis.set_minor_locator(AutoMinorLocator(n=n_xminor))\n",
    "    ax.yaxis.set_minor_locator(AutoMinorLocator(n=2))\n",
    "\n",
    "    if fig is not None and save:\n",
    "        export_file = plots_dir.joinpath('hist_{}.png'.format(f_name.lower()))\n",
    "        fig.savefig(export_file, bbox_inches='tight')\n",
    "        print(export_file)\n",
    "\n",
    "        export_file = export_file.with_suffix('.pdf')\n",
    "        fig.savefig(export_file, bbox_inches='tight')\n",
    "        print(export_file)\n",
    "\n",
    "    return fig, ax\n",
    "\n",
    "# # Try.\n",
    "# f_name = 'num_attempt'\n",
    "# step = 1\n",
    "# bins = np.arange(0 - step / 2, 20 + step / 2, step)\n",
    "# fig, ax = plot_hist(f_name, bins, legend=False)\n",
    "# ax.set_xlabel('number of attempts')\n",
    "# ax.set_xlim([-0.5, 21])\n",
    "# ax.set_xticks(np.arange(0, 20 + 1, 4))\n",
    "# ax.set_ylim([0, 10.5])\n",
    "# ax.set_yticks(np.arange(0, 8.1, 4));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set font sizes.\n",
    "plt.rc('font', size=36)  # controls default text sizes\n",
    "plt.rc('axes', titlesize=36)  # fontsize of the axes title\n",
    "plt.rc('axes', labelsize=36)  # fontsize of the x and y labels\n",
    "plt.rc('xtick', labelsize=36)  # fontsize of the tick labels\n",
    "plt.rc('ytick', labelsize=36)  # fontsize of the tick labels\n",
    "plt.rc('legend', fontsize=22)  # legend fontsize\n",
    "plt.rc('figure', titlesize=40)  # fontsize of the figure title"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "../figures/attempt_and_turn_histograms.pdf\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, 2, sharey=True, figsize=(16, 4))\n",
    "\n",
    "ax = axes[0]\n",
    "f_name = 'num_attempt'\n",
    "step = 1\n",
    "bins = np.arange(0 - step / 2, 20 + step / 2, step)\n",
    "plot_hist(f_name, bins, legend=True, ax=ax)\n",
    "ax.set_xlabel('number of attempts')\n",
    "ax.set_xlim([-0.5, 21])\n",
    "ax.set_xticks(np.arange(0, 20+1, 4))\n",
    "ax.set_ylim([0, 10.5])\n",
    "ax.set_yticks(np.arange(0, 8.1, 4))\n",
    "ax = axes[1]\n",
    "f_name = 'num_turn'\n",
    "step = 5\n",
    "bmax = 100\n",
    "bins = np.arange(0, bmax + step, step)\n",
    "plot_hist(f_name, bins, legend=False, ax=ax)\n",
    "ax.set_xlabel('number of turns')\n",
    "ax.set_xlim([20 - 0.3, bmax + 0.3])\n",
    "ax.set_xticks(np.arange(20, bmax+1, 4*step))\n",
    "ax.set_ylim([0, 10.5])\n",
    "ax.set_ylabel('')\n",
    "ax.set_yticks(np.arange(0, 8.1, 4))\n",
    "\n",
    "\n",
    "fig.tight_layout(pad=0.4, w_pad=1.5, h_pad=1.0)\n",
    "\n",
    "export_file = plots_dir.joinpath('attempt_and_turn_histograms.png')\n",
    "# # Export png (disabled)\n",
    "# fig.savefig(export_file, bbox_inches='tight')\n",
    "# print(export_file)\n",
    "\n",
    "export_file = export_file.with_suffix('.pdf')\n",
    "fig.savefig(export_file, bbox_inches='tight')\n",
    "print(export_file)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (JUSThink-alignment)",
   "language": "python",
   "name": "justhink-alignment-env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}